The present invention relates to a network, and particularly, to a network that extends among a plurality of domains.
In a network, domains managed by different network management servers or a domain (for example, an autonomous system (AS) on the Internet) controlled according to one policy is present. When a network service (hereinafter, referred to as an interdomain service) that extends among the plurality of domains intends to be developed, in the case where system models of both parts correspond to each other and control and management information for each of the models is interconverted to perform setting, controlling, and management throughout the plurality of domains, it is considered that setting, controlling, and the management can be more easily performed than a case where the setting, controlling, and management cannot be performed.
The control and management information needs to be transmitted and received among domains in order to implement the interdomain service, but methods therefor include a centralized management method of performing management by using a small number of mechanisms such as a network management server as above and a distributive control method of performing a control by transmitting and receiving a control message among general nodes or control nodes which are distributed in the domains.
(Federation on Centralized Management)
Therebetween, in regard to a setting and management method of the interdomain service by the centralized management method, an example of applying a comparatively loose centralized control method to a virtual network (hereinafter, referred to as a slice) is disclosed in US Patent Application Publication No. 2011/0149968, “Method for Controlling Internet Network”. In US Patent Application Publication No. 2011/0149968, “Method for Controlling Internet Network”, a network management server (clearing house, aggregate manager) performs only issuing of an identifier of the slice and authentication and authorization and an authorized user (service provider) directly performs setting or management operations of resources of each domain by using an access key acquired from the network management server. In US Patent Application Publication No. 2011/0149968, “Method for Controlling Internet Network”, each network management server handles only a resource enclosed in a management region thereof and does not handle resources that extends among a plurality of regions or a plurality of management domains like network link resources (virtual link resources).
Further, “Development Method and Experience of A Non-IP Protocol Using the Virtualization Nodes” written by Kanada Yasusi and Nakao Akihiro, IEICE Technical Committee on Internet Architecture, Phrases 3.2, 4.1, and 4.2 of Chapter 2, 2010-12-17 discloses an example of adopting the centralized management method which is more centralized. In “Development Method and Experience of A Non-IP Protocol Using the Virtualization Nodes” written by Kanada Yasusi and Nakao Akihiro, IEICE Technical Committee on Internet Architecture, Phrases 3.2, 4.1, and 4.2 of Chapter 2, 2010-12-17, a domain controller (network management server) serves to generate a slice, that is, a virtual network on a physical network. That is, the domain controller receives a slice definition including components of the slice, that is, a node sliver (virtual node), a link sliver (virtual link), and a slice definition including combination information of the node sliver and the link sliver which are included in the slice from a developer of the slice, and distributes the slice definition to a virtualization node (physical network node). Each virtualization node allocates necessary resources to the node sliver and the link sliver according to the slice definition.
(Federation on Distributive Control)
Meanwhile, an example of setting and management method of the interdomain service by the distributive control method is disclosed in Private Network-Network Interface Specification Version 1.0 (PNNI 1.0), af-pnni-0055.000, ATM Forum, March 1996, p. 13 to p. 38. In Private Network-Network Interface Specification Version 1.0 (PNNI 1.0), af-pnni-0055.000, ATM Forum, March 1996, p. 13 to p. 38, a peer group (domain) including a plurality of physical nodes is handled as a representative node by using a private network-network interface (PNNI) to simplify routing information (control information) and configure a network which can be scalably routed (controlled). In Private Network-Network Interface Specification Version 1.0 (PNNI 1.0), af-pnni-0055.000, ATM Forum, March 1996, p. 13 to p. 38, the peer group including a plurality of peer groups is hierarchically configured to implement new scalability.
(Scalability)
In most domains by the centralized management method, since the network management server directly manages contents or the amount of resources as well as the network management server manages only an access authority as in US Patent Application Publication No. 2011/0149968, “Method for Controlling Internet Network”, it is necessary to access the resources through the network management server at all times. In order to implement unitary setting and management throughout the other domain in the domain, the network management server needs to distinguish management information of an own domain from management information of the other domain and use different formats in a message within the domain and a message outside the domain. Further, by performing the distinguishment, scalability regarding a manageable network scale is ensured.
This point is similarly applied even to negotiation of control information between control nodes in the distributive control method. For example, on the Internet, in routing in a domain (Autonomous System (AS)), an Interior Gateway Protocol (IGP) such as a Routing Information Protocol (RIP), Open Shortest Path First (OSPF), or the like is used, and in interdomain routing, an Exterior Gateway Protocol (EGP) such as a Border Gateway Protocol (BGP), or the like is used. Further, a PNNI is used even with respect to a peer group of any hierarchy in an ATM, but there is a difference in that a message in the peer group includes database information within the peer group and the message outside the peer group does not include the database information within the peer group. That is, the scalability is ensured by distinguishing and using an IGP and an EGP within and outside the domain in the Internet. Meanwhile, database information regarding network topology, and the like is transmitted to the inside of the peer group in the ATM, but the database information is not transmitted to the outside of the peer group to ensure the scalability. In “Using events for the scalable federation of heterogeneous components” written by J. Bates, J. Bacon, K. Moody, and M. Spiteri, Proceedings of the 8th ACM SIGOPS European workshop on Support for composing distributed applications, pp. 58-65, 1998, a network is assumed, in which all of respective domains adopting the EGP use the Internet protocol and further, in Private Network-Network Interface Specification Version 1.0 (PNNI 1.0), af-pnni-0055.000, ATM Forum, March 1996, p. 13 to p. 38, it is assumed that all of the peer groups use the ATM and the PNNI.
(Federation on Event Base)
“Using events for the scalable federation of heterogeneous components” written by J. Bates, J. Bacon, K. Moody, and M. Spiteri, Proceedings of the 8th ACM SIGOPS European workshop on Support for composing distributed applications, pp. 58-65, 1998 discloses a method of performing federation among a plurality of domains by using the event base. In this method, the element (management server) of each domain is the event base and for example, it is assumed that this method is operated according to the API such as event notification, event notification reservation, and the like. In this method, in addition of a federation function, a system federator linking the plurality of domains is added and an operation to be made when the system federator receives event notification as well as the event is reserved to be notified to the system federator when a predetermined event is generated in each domain is described as a rule. This method cannot be applied when each domain does not have an operating mechanism of the event base.